Enhancing Cycling Stability of Aqueous ... - ACS Publications

Research Article www.acsami.org

Enhancing Cycling Stability of Aqueous Polyaniline Electrochemical Capacitors Luciano M. Santino,† Yang Lu,‡ Shinjita Acharya,† Liana Bloom,† Daniel Cotton,† Aly Wayne,† and Julio M. D’Arcy*,†,‡ †

Department of Chemistry and ‡Institute of Materials Science & Engineering, Washington University in St. Louis, St. Louis, Missouri 63130, United States S Supporting Information *

ABSTRACT: Electrochemical capacitors fabricated with polyaniline nanofibers are cycled 150 000 times with 98% capacitance retention. These devices maintain an energy density of 11.41 Wh/kg at a power density of 4000 W/kg, 64 times greater than that of an identically fabricated device based on activated carbon (0.177 Wh/kg at 4600 W/kg). For applications requiring a higher specific energy, 33.39 Wh/kg at a specific power of 600 W/kg is obtained by widening the voltage window; this device retains 93% capacitance after 10 000 cycles. We achieve a high cycling stability through careful device engineering paired with a renewed focus on the electrochemical processes occurring at the positive and negative electrodes during cycling. KEYWORDS: PANi, conducting polymers, energy storage, pseudocapacitance, doping, EDLC, supercapacitors, poly(lactic acid)

■

and discharging. Covalent or metallic bonding in solids of fixed crystalline structure facilitates ion and electron transport;4 however, potential induced phase changes can cause cracking, leading to dramatic rises in electrode resistance and a rapid loss of cycling stability.5 Furthermore, low electrical conductivity of metal oxides and scarcity of some transition metals motivates the engineering of devices utilizing organic semiconductors. Conducting polymers, whose development began in the early 1960s,6−9 are ideal candidates for pseudocapacitive energy storage materials due to their low cost, facile nanostructure synthesis, processability, and high energy density at high power density. Conducting polymer nanofibers are van der Waals bound bundles of rodlike conjugated chains of carbon atoms10 that, unlike carbon or metal oxides, are viscoelastic, allowing for structural changes without mechanical breakdown.11 There are several complementary figures of merit utilized to optimize energy storage devices. At a base level, the total amount of charge (q) stored by a particular capacitive material should be increased. This can be estimated through cyclic voltammetry and eq 1:

INTRODUCTION World consumption of renewable energy is expected to increase by 2.6% every year over a projected period between 2012 and 2040.1 As renewables are intermittent sources of power, energy storage and grid stabilization are booming fields of research. Electrochemical capacitors (ECs), colloquially known as supercapacitors, are a family of energy storage devices with moderate energy density at high power densities thus delivering a large amount of charge in a short time. ECs differ from batteries because relevant charge storing events occur mostly on the surface rather than in the bulk of the material, leading to a decrease in energy density concomitant with an increase in power density.2 Activated carbon is the prototypical material for commercial electric double layer capacitors (EDLCs). Activated carbon relies on superficial double-layer capacitance,3 a mechanism that unfortunately limits the achievement of a high energy density. Significant effort in recent years aims to improve existing supercapacitor technology using materials that store energy through doping or superficial charge-transfer reactions, i.e. through reduction and oxidation, a mechanism known as pseudocapacitance. Increasing energy density using pseudocapacitors is catalyzing development of the next generation of energy storage devices. Realization of this impactful vision depends on engineering stable pseudocapacitors of high cyclability. Metal oxides and conducting polymers constitute the stateof-the-art in pseudocapacitive materials. Energy density in metal oxides is high due to redox reactions occurring during charging © 2016 American Chemical Society

q=

∫0

t

I dt

(1)

where 0 and t mark the beginning and end of a discharge cycle, respectively. Increasing q while maintaining a constant Received: August 5, 2016 Accepted: October 10, 2016 Published: October 20, 2016 29452

DOI: 10.1021/acsami.6b09779 ACS Appl. Mater. Interfaces 2016, 8, 29452−29460

Research Article

ACS Applied Materials & Interfaces

to the kinetically facile and highly reversible protonic doping and dedoping processes of the emeraldine base and salt states of polyaniline. Thus, a PANi device, even while constrained to a smaller voltage window, exhibits a moderate specific energy with remarkable power and cycling characteristics. It is also possible to expand our EC’s voltage window to 0.8 V, increasing specific energy to 33.39 Wh/kg at 600 W/kg. By first cycling a freshly fabricated device at 0.5 V, a PANi-based electrode architecture experiences less of a strain when voltage is increased to 0.8 V. This initial priming step allows our device to retain 93% capacitance after 10 000 cycles under 0.8 V at 10 A/g, as opposed to only 87% without any priming. Structural changes occurring when a PANi device is cycled to 0.8 V cause breakdown of the percolation network, disadhesion from the current collector, and cycling instability.23 Priming relaxes the molecular structure of the nanofibers graduallyimproving their electrochemical stability under higher voltage windows. PANi is synthesized through solution-based dimer initiated oxidative polymerization. The colloidal dispersion is then filtered and washed with deionized water before being lyophilized while still wet. Figure 1 shows the morphology of

electrode mass, i.e. by increasing the accessible surface area of a material, leads to an increase in its specific capacitance, Csp (eq 2): q Csp = (2) mΔV where m and ΔV represent the mass and absolute voltage window, respectively. Specific capacitance is a reliable metric for comparison between electrodes studied under similar electrochemical conditions. However, it is important to normalize for ΔV in order to determine the applications for a particular material. An EC with ΔV = 0.1 V, for example, could have a high Csp, but would not be able to power an LED at 1.5 V unless connected in series with 15 other ECs. Specific energy, U, is a measure of the total electrochemical energy stored by a cell, and is proportional to ΔV2 (eq 3).

U=

1 CspΔV 2 2

(3)

The total stable voltage window of a device is constrained by a myriad of factors. Aqueous electrolytes are theoretically limited to a maximum voltage window of 1.23 V due to the electrolysis of water at the positive and negative electrodes. In practice, symmetric aqueous ECs are limited to less than 1 V to maximize device lifetime.12 Some organic electrolyte based devices, including EDLCs, can operate with potentials as high as 3 V. Given two devices with identical activated carbon electrodes in an organic electrolyte, a device charged to 3 V will store 9 times as much energy as a device at 1 V based on voltage window alone. However, designing conducting polymer electrochemical capacitors with organic electrolytes is not simple, and often leads to poor cycling stability with little or no improvement in energy density.13−15 As ECs are intended to complement batteries in applications where rapid, stable cyclability at a moderate energy density is a necessity, cycling stability is the foremost metric determining their success. Herein, we report an extraordinarily stable aqueous electrochemical capacitor based on polyaniline nanofibers (PANi). This device retains 98% of its initial capacitance after 150 000 charge−discharge cycles at 25 A/g, a vastly improved cycling stability rivalling that of state-of-the-art aqueous carbon-based EDLCs. This performance enhancement is achieved through a focus on the electrochemical processes occurring in a PANi electrode during charging and discharging, as well as carefully choosing compatible device components.

■

Figure 1. Micrographs of PANi: (a) SEM of pristine purified PANi; (b) TEM of pristine purified PANi; (c) SEM of poly(lactic acid) web cast on hard carbon paper magnified to 7000× and (d) 30000×.

RESULTS AND DISCUSSION Often, the well-studied “swelling” due to ion intercalation between polymer chains16,17 is noted as the primary reason for the loss of capacitance and increase in resistivity when cycling conducting polymer electrodes.18,19 The previous highest reported cycling stability for a polypyrrole thin film (1 μm active material thickness) EC shows 80% retention after 130 000 cycles at 3 A/g for a 0.6 V window.20 A contributing factor in the stability of this device is the researchers’ use of a constrained voltage window. For a material like activated carbon where capacitance is inherently low, the voltage window must be maximized in order to increase energy density (eq 3); in conducting polymers, this is not the case. Our device exhibits a specific energy of 11.41 Wh/kg at a specific power of 4000 W/kg at 0.5 V, nearly 65 times greater than that of an identically fabricated device based on activated carbon at 0.8 V (0.177 Wh/kg at 4600 W/kg). We attribute this performance

PANi as cast from an aqueous colloidal suspension (Figure 1a) purified through washing with chloroform and deionized water, as well as from a dispersion of lyophilized PANi in hexafluoroisopropanol (Figure 1b). The PANi is then dispersed into a solution containing 2% poly(lactic acid) in benzene. The biopolymer poly(lactic acid), PLA, aids in the dispersion of PANi in many solvents, and doubles as a binding material for stable electrode casting. Figure 1c,d shows scanning electron micrographs (SEM) of the electrode active material’s unique microarchitecture. When a PANi/PLA/benzene slurry dries under directed humid airflow, droplets of water condense on the surface of the rapidly evaporating mixture, resulting in a PLA web embedded with PANi.24−27 This web covers a nanofibrillar core extending to the hard carbon paper current collector, leading to an excellent power response without the 29453


Research Article


appear during cycling past 0.6 V (Figure S1) and continue to grow with symmetric and closely associated reduction peaks. Previous studies have attributed these peaks to the degradation of polyaniline to soluble redox pairs such as benzoquinone and hydroquinone and p-aminophenol and benzoquinoneimine.32−34 Their current density grows throughout cycling, suggesting an irreversible process which may be deleterious to cycling stability. The reduction peak at a cathodic 0.07 V exhibits no change in current density, suggesting that it is independent of the electrode reduction occurring below 0 V. Constraining the cycling potential between 0 and 0.5 V vs Ag/ AgCl leads to a stable and reversible cyclic voltammogram. The cathodic peak at 0.07 V and the anodic peak at 0.21 V vs Ag/ AgCl are still present, which could be contributions from a highly reversible leucoemeraldine/emeraldine redox cycle. In the charging direction, PANi is p-doped to a potential of 0.80 V vs Ag/AgCl, bringing it to its emeraldine salt form. Discharging to 0 V vs Ag/AgCl fully dedopes PANi to emeraldine base (Figure 2b). This charging−discharging process can repeat reversibly over many cycles with minimal capacitance loss.35−39 UV−vis spectra show that a colloidal dispersion of PANi containing aqueous HCl expresses a freecarrier tail, suggesting that it is in its electronically conductive doped emeraldine salt state. Adding aqueous ammonia to the dispersion results in a drastic change in color to blue, and a loss of the free-carrier tail, suggesting dedoping to emeraldine base (Figure 2c). As doping should lead to an increase in conductivity in the solid state, a film of PANi was cast onto glass and exposed to HCl and NH3 vapors. An I−V curve taken of the films shows that the dedoped film (blue) is highly resistive when compared to the doped film (green). Three-electrode cyclic voltammetry helps predict the chemical events occurring at the anode and cathode in a twoelectrode experiment. In a symmetric PANi EC, both electrodes begin at the same open circuit potential (OCP) of 0.44 V vs Ag/AgCl, leading to a net potential between the two electrodes of 0 V. In practice, the OCP of our devices is 0.0021 ± 0.0071 V. This difference stems primarily from nonunity mass ratios and imperfect area of overlap of the electrodes. Upon charging, the negative electrode is dedoped from its initial state and the positive is further p-doped. Once the negative electrode is fully dedoped, further increasing the polarization of the device causes the electrochemical reduction of the negative electrode to leucoemeraldine, an insulator,40 while the positive electrode remains in its emeraldine salt form. Thus, while a certain PANi electrode may appear to have a high cycling stability with a 0.8 V window in a three-electrode geometry, an EC made with two PANi electrodes often shows a low cycling stability for similar potential windows.41,42 Figure 3 details our fabrication process. The stability and energy density of an EC is highly dependent on the choice of electrolyte and current collector. Sulfuric acid (1 M) is used as an electrolyte due to its high ionic conductivity, ideal for studying PANi’s protonic doping mechanism (Figure S2). Copper and aluminum are the archetypal current collectors for lithium ion batteries because both metals are abundant and stable within their respective potential regions in organic electrolytes. Unfortunately, when applied to aqueous systems, copper is readily oxidized to Cu2+ and aluminum is oxidized to Al3+. In our aqueous system, hard carbon paper (Figure S3) serves as a stable current collector. Hard carbon paper is mostly graphitic and exhibits an insignificant capacitance when cycled in sulfuric acid (Figure S4). The Pourbaix diagram in Figure 4

addition of carbonaceous conductors such as acetylene black or carbon nanotubes. The active material in this study extends 200−300 μm above the surface of the hard carbon paper, well within the reported range of thickness for a commercial EC electrode.28 Three-electrode cyclic voltammetry aids in determining a stable voltage window for PANi immersed in electrolyte. Figure 2a shows a three-electrode cyclic voltammogram of a PANi

Figure 2. Characterization of PANi. (a) Three-electrode cyclic voltammogram of PANi/PLA on hard carbon paper in 1 M H2SO4 at 25 mV/s. Note: The 0−0.5 V cycle is for a freshly made electrode with a lower mass. (b) Within the window of 0−0.5 V vs Ag/AgCl in an acid, PANi is doped to emeraldine salt and dedoped to emeraldine base. (c) UV−vis spectra of a purified colloidal dispersion of PANi in water, as doped with aqueous HCl and dedoped with aqueous NH3. (d) I−V curves of PANi/PLA films, exposed to HCl and NH3 vapors, respectively. The doped film shows a vastly increased conductivity over the dedoped film. Inset: Photographs of colloidal dispersions of PANi: doped and dedoped.

electrode in 1 M H2SO4. An electrode fabricated from pristine PANi cast on hard carbon paper without PLA exhibits an open circuit potential of 0.37 V vs Ag/AgCl in 1 M H2SO4, within range of a partially doped emeraldine salt form of polyaniline.29 When cast from a benzene solution containing PLA, the open circuit potential is 0.44 V, suggesting further doping of PANi by PLA. Polyaniline has insulating reduced (leucoemeraldine, 0.07 V, cathodic) and oxidized (pernigraniline, 0.80 V, anodic) states that lead to the low current (flat regions) at both low and high potentials. The insulating state is localized around the current collector/PANi interface, where the effective current density experienced by the PANi active material is the greatest. The formation of such an insulating layer, with a thickness dependent on the amount of charge transferred, i.e. time spent at those potentials, acts as an open “switch” between the current collector and the bulk of the active material,30 leading to a quick loss of charge density in PANi. An applied potential closes the switch by oxidizing or reducing the insulating layer back to a conducting state.31 The intensity of the chargetransfer peaks is dependent on reactions occurring outside of PANi’s fingerprint region, i.e., below 0.07 V and above 0.80 V vs Ag/AgCl. The oxidation peaks around 0.50 and 0.57 V 29454


Research Article


permeability of the separator to ions and S is the area of the separator.46 Careful mass-balancing of electrodes leads to extended cycling stability47 and allows for more accurate measurements in large-scale devices.48 For the purposes of this work, when M E− < 1, the positive electrode, E+, is more massive than the + ME

negative electrode, E−. When

M E− M E+ −

> 1, E− is more massive than

E+. In Figure 5, the mass of E is greater than the mass of E+ by

Figure 3. Fabrication process of electrochemical capacitors. (a) A platinum foil lead is affixed to a piece of polyimide tape. (b) Hard carbon paper current collector coated with drop-cast active material is placed on platinum. (c) Celgard 3501 serves as an ion-permeable separator to prevent a short circuit. (d) Opposite electrode is placed with active material facing the separator. (e) A platinum lead rests on the electrode. (f) The device is sealed using a piece of polyimide tape. (g) The supercapacitor is coated with 5 min epoxy for long-term cycling ability.

Figure 4. Pourbaix diagrams of common current collectors. These pH potential diagrams show (a) stability of water with different pHs under applied potentials vs SHE. Note: To convert to E vs Ag/AgCl, subtract 0.209 V at 25 °C. (b) Stability windows for platinum, graphite, aluminum, and copper in water.

Figure 5. Electrochemical properties of ECs where

M E− M E+

= 1.25. (a)

Cyclic voltammograms recorded at 25 mV/s. Red cycles are for a 0.8 V window; black cycles are for a 0.5 V window. Solid lines represent the second cycle of the cyclic voltammogram and dashed lines represent the 25th. (b) A Nyquist plot shows that devices charged to 0.5 V exhibit less charge-transfer resistance than devices charged to 0.8 V. (c) When a symmetric supercapacitor is cycled to 0.8 V at 5 A/g, it drops to 20% of its initial capacitance over 10 000 cycles. A similar device constrained to 0.5 V shows marginal capacitance loss over 7000 cycles due to electrolyte evaporation. The 0.5 V supercapacitor is stable for >150 000 cycles at 25 A/g when sealed in epoxy.

shows that graphite degrades at most potentials and pH into more thermodynamically stable materials such as methane and CO2.43 However, these reactions necessitate a four-electron transfer, presenting an insurmountable kinetic barrier.44 Graphitic carbon therefore proves to be a suitable, costeffective current collector for aqueous acidic electrolytes. As such, one attempt to improve the stability of aqueous lithium ion batteries involves coating electrode materials with carbon to inhibit corrosion.45 Casting PANi/PLA films onto hard carbon paper also enhances adhesion of the conducting polymer/ binder mixture vs casting directly onto untreated metallic substrates such as platinum. Increased adhesion helps mitigate the cycling stability loss caused by any morphological changes of the polymer network upon charging and discharging. Platinum foil serves as a lead for electrodes, because it enables well-sealed devices thus reducing the contribution of electrolyte evaporation during long cycling stability tests. The entire EC can be encased in an epoxy resin to further mitigate evaporative processes. Celgard 3501 is chosen as the separator as it is wettable by 1 M H2SO4 and only 25 μm thick, whereas glass fiber paper and filter paper are ∼300 and ∼200 μm thick, respectively. The use of a thinner separator can lead to a lower equivalent series resistance (ESR) of the packaged device. The resistance of the separator, RS = ϱSδSS−1, increases with separator thickness, δS. In this equation, ϱS refers to the

25% ( M E+ = 1.25). The potentials reported describe the −

ME

potential of the positive electrode with respect to the negative electrode. When this device is cycled between 0 and 0.5 V, its cyclic voltammogram remains stable over the first 25 cycles (Figure 5a, black). Another EC with the same mass balancing is cycled between 0 and 0.8 V; its cyclic voltammogram changes from one large truncated peak at 0.8 V to two pairs of peaks centered around 0.3 V over only 25 cycles (Figure 5a, red). To rationalize this, we must consider the electrochemical processes occurring at each electrode. As previously stated, the two electrodes begin at the same open circuit potential (0.44 V vs Ag/AgCl) when freshly fabricated; the device’s total OCP is ∼0 V.49 Initial cycling causes simultaneous dedoping of the negative electrode and p-doping of the positive electrode; throughout this process, each electrode reaches a stable resting potential. Unfortunately, it is difficult to directly correlate the 29455


Research Article


electrodes, the voltage window of an electrochemical capacitor can be extended. This configuration is known as a type II conducting polymer EC, or an asymmetric supercapacitor (ASC).53 Much of the recent success of ASCs comes from increased focus on the stable potential range of electrode materials; interestingly, there are many similarities between symmetric and asymmetric ECs. Both electrodes in an EC cell experience different potential ranges during charging and discharging, regardless of whether or not they are made from the same active material; the kinetics of doping and dedoping do not mirror each other. It is thus imperative to understand the range of stability of a material, even for symmetric devices; otherwise cycling stability will be adversely affected. One limitation of ASCs is that they are often fabricated with neutral electrolytes to ensure reasonable cycling stability. A 1 M H2SO4 solution with a low pH has a high concentration of H+, reducing the overpotential for hydrogen evolution, making the negative electrode unstable in a device (see Figure 4). Gel electrolytes may help mitigate this effect, as an asymmetric (−)PEDOT/PANi(+) electrochemical capacitor cycled in an acidic gel electrolyte with a total potential window of 1.6 V shows around 80% capacitance retention over 10 000 cycles.54 Instability associated with gas evolution often outweighs any increased energy density afforded by a wider voltage window, apart from devices comprised of high surface area carbon electrodes. The strong adsorption of hydrogen in the pores of activated carbon, for example, creates a sizable overpotential for H2 evolution that allows activated carbon electrodes to be cycled down to −1 V vs Ag/AgCl in 1 M H2SO4.55 This trait makes activated carbon popular as a negative electrode material for ASCs cycled in acidic electrolytes. However, simply using an activated carbon negative electrode with a conducting polymer or metal oxide positive electrode is not an effective solution to the cycling stability issues associated with pseudocapacitive materials. A recent example shows an (−)activated carbon/ PANi(+) ASC cycled to a high voltage window of 1.4 V in 1 M H2SO4. While this device appears stable over 2000 cycles,38 its specific energy is only 14.1 Wh/kg at a specific power of 70 W/ kg. At only 0.8 V, the all-PANi symmetric device presented in our work gives a much higher specific energy of 33.39 Wh/kg at a specific power of 600 W/kg. To determine the effect of mass loading on the electrical properties of each electrode, two-point probe resistance values were collected on multiple samples after cycling to 0.8 V. The control χ, a PANi film cast on a hard carbon paper substrate soaked in 1 M H2SO4 for 1 week, shows a resistance of 164.8 ± − 23.7 Ω (Figure 6a). A device i was fabricated where M E+ = 0.5,

peaks in a two-electrode cyclic voltammogram to those of a three-electrode cyclic voltammogram as in Figure 2, as the peaks of the negative and positive electrodes are superimposed. A powerful tool known as electrochemical impedance spectroscopy (EIS) allows us to distinguish between kinetically limited charge-transfer events and capacitive behavior.50 Newly packaged devices are cycled 100 times within voltage windows of 0.5 and 0.8 V, respectively. The fully charged potential is then held for 30 min to achieve a steady state, at which point an ac signal of 10 mV is applied and the ac response is recorded over frequencies ranging from 200 kHz to 15 mHz. A Nyquist plot (Figure 5b) shows the imaginary impedance −Z″ plotted against the real impedance Z′. By examining the x-intercept of the Nyquist plot, the equivalent series resistance (ESR) of the packaged device is found to be 1.500 Ω when charged to 0.8 V and 0.972 Ω when charged to 0.5 V. The characteristic semicircular region apparent at high frequencies for the 0.8 V device suggests a large charge-transfer resistance brought on by kinetically slow redox events. This is in contrast to the 0.5 V device, which shows no such semicircle, appearing capacitive in nature. We attribute this phenomenon to the facile kinetics of protonically doping and dedoping PANi from its emeraldine salt to emeraldine base form over the arduous redox events (leucoemeraldine ← emeraldine → pernigraniline) occurring at extended voltage windows. The stability of the EC over many charge−discharge cycles is determined by plotting capacitance retention vs number of cycles (Figure 5c). The specific capacitance of an EC is most realistically estimated from galvanostatic charge−discharge experiments. Specifically, the applied current is divided by the slope of the discharge curve and mean electrode mass m (eq 4).51 ⎛ dV ⎞−1 Csp = I ⎜ ⎟ m−1 ⎝ dt ⎠

(4)

The discharge curve is utilized rather than the charge curve, as some charge is lost due to Coulombic inefficiency; i.e., calculation based on the charging curve overestimates the capacitance. The 0.5 V device is stable for >150 000 cycles at a high rate of 25 A/g with 98% capacitance retention. This is comparable to the cycling stability of EDLCs in aqueous electrolytes, albeit under a slightly constrained voltage window. When the device is cycled to 0.8 V, capacitance drops rapidly. Cycling stability is often defined as the number of charge− discharge cycles it takes for a device to lose 20% of its initial capacitance,52 which is around 2200 cycles for an EC with a more massive negative electrode cycled to 0.8 V. There is a clear discrepancy in cycling stability between − devices cycled to 0.8 and 0.5 V. When M E+ = 1.25, the increased

ME

and cycled to 0.8 V for 5000 times at 10 A/g. The lack of discernible change in resistance from χ suggests that the positive electrode’s excess mass kept the negative electrode − from experiencing reducing potentials. Device ii, where M E+ =

ME

mass ratio of the negative electrode over the positive electrode shifts the overall cell polarization to more negative potentials. When the device is cycled to only 0.5 V, the shift in polarization is not large enough to form an insulating state at either electrode. In a window spanning 0.8 V, the negative electrode experiences a low enough potential to be reduced from emeraldine to leucoemeraldine. This dramatically increases the resistance at the current collector/PANi interface, leading to loss of device power capability, and a significant loss of device stability. There are many reported methods of increasing conducting polymer EC cycling stability. By employing materials with complementary voltage windows for the positive and negative

ME

1.45, was also cycled to 0.8 V for 5000 times at 10 A/g, and shows a dramatic increase in resistance of the negative electrode to 757.0 ± 60.9 Ω. Figure 6b shows the current−potential (I− V) curve characteristics of both device i and device ii vs control χ. The negative electrode of device ii shows a deviation from the normal linear response associated with an ohmic material, implying reduction to leucoemeraldine. Noticeably, the positive electrode remains similar in resistance to χ, indicating that the potential window is not sufficiently anodic to oxidize emeraldine to pernigraniline. 29456


Research Article


Figure 6. I−V and XPS characterization of electrodes after cycling, − − where in i M E = 0.5, in ii M E = 1.45, and χ is an uncycled control. (a) M E+

Figure 7. Characterization of devices where voltammogram of a device where

M E+

Resistance values for positive and negative electrodes of three devices. (b) I−V curves for i show electrodes exhibit a similar curve and ii’s negative electrodes exhibit nonlinear behavior after cycling. (c) High resolution C1s spectra of i, ii, and χ.

M E− M E+

M E− M E+

< 1. (a) Cyclic

= 0.98 at 25 mV/s, constrained

to a ΔV of 0.5 V for 10 000 cycles at 25 A/g and then cycled to 0.8 V − at 25 mV/s. (b) Cycling stability for an EC where M E = 0.82 encased M E+

in epoxy initially cycled to 0.5 V at 10 A/g and then to 0.8 V after 7000 cycles. (c) Cycling stability for devices initially primed using 0.5 V for 10 000 cycles at 25 A/g and then switched to a wider window of 0.8 V − at 10 A/g, and for a device cycled solely at 0.8 V using 10 A/g; M E =

X-ray photoelectron spectroscopy (XPS) probes changes in the chemical bonding environment of PANi. Figure 6c shows the high resolution C1s spectra of χ, and of the electrodes in devices i and ii. The C1s spectra is deconvoluted in three components (Figure S5). These peaks are assigned to C−C or C−H of the aromatic ring at ∼283 eV, C−N or CN at ∼284.7 eV, and C−N+, CN+, or CO at around 287 eV.56 Though the overall shape remains unchanged throughout the electrodes, there is a small difference in the intensity ratio of the peak centered at ∼284.7 eV. Qualitatively, device i has almost identical peak ratios to χ, indicating retention of the original chemical composition of the active material. However, for device ii, the peak at 284.7 eV is broadened and of lower intensity, possibly a result of new bonding interactions under anodic and cathodic cycles. XPS spectra were collected at the PANi/electrolyte interface, and this technique probes the top 10 nm of the electrode surface;57 changes in PANi composition are possibly more prevalent at the current PANi/current collector interface. Clearly, the negative electrode’s mass affects the electrochemical stability of the EC in a 0.8 V window. Figure 7a shows − the cyclic voltammogram of a device where M E+ = 0.98. For a

M E+

0.8 for both devices.

over 10 000 cycles. The difference in capacitance retention between these two devices is readily apparent over the first few thousand cycles. Initial cycling at 0.5 V slows the initial capacitance fade of a device by gently coaxing the electrolyte ions into the bulk of the active material. Cycling directly at 0.8 V, without priming, leads to fast capacitance loss due to the rapid increase in turgidity causing structural breaks in the percolation network of the PANi electrodes. To further compare devices against one another, normalized values of specific energy are plotted against specific power in what is known as a Ragone plot. Electrochemical capacitors show a characteristic curve displaying the dependence of specific energy (U) on specific power (P). A lower P will allow more time for slow electrochemical processes to occur, leading to a higher U. An electrolytic capacitor tends to show little to no dependence of U on P, and a battery shows a high U at low P, dropping off rapidly at higher P. Specific power is calculated using eq 5:

ME

0.5 V window, two small pairs of peaks are present due to protonic doping or catalyzed oxygen side reactions (see Figure 2b).22 Initially, the device is cycled to 0.5 V for 10 000 times at 25 A/g via galvanostatic charge−discharge. The voltage window is then extended to 0.8 V and the device is cycled 25 times at 25 mV/s; notably, the cyclic voltammogram appears similar to that of Figure 5a. Two differences are an increased switching rate and the absence of a broad peak at 0.8 V, attributable to the suppression of reduction processes at the negative electrode. The current density ratios of the peaks at 0.3 and 0.4 V are also switched from that of Figure 5a. Figure 7b shows the cycling stability of an EC initially cycled to 0.5 V for 7000 cycles, and then cycled to 0.8 V. A jump in capacitance occurs upon window expansion. The capacitance then rapidly drops and stabilizes to a value higher than that of − the 0.5 V window. Figure 7c shows two devices where M E+ =

3600 CspV 2t −1 (5) 2 where t is the total time of discharge for one cycle and Csp is calculated from eq 4 for multiple applied current densities. In Figure 8a the Ragone plot of four ECs fabricated using identical device architectures (Figure 3) shows the similarity between unprimed (i) and primed (ii) devices. While a primed device shows a higher cycling stability as evidenced by Figure 7c, its Ragone curve is nearly identical to that of an unprimed device; in other words, priming does not affect the amount of active material involved in charge storage. Rather, the polymer nanofibers at each electrode relax in the electrolyte over the course of priming, leading to less strain when the voltage window is expanded to 0.8 V and swelling occurs. Figure 8b shows the galvanostatic charge−discharge curves at 0.5 and 1 A/g in a 0.8 V window for a PANi EC versus an activated carbon EC. Because capacitance is proportional to the length of discharge, Figure 8b shows that PANi devices are characterized by a higher specific capacitance than activated carbon devices (100.8 F/g at 0.54 A/g vs 15.66 F/g at 0.16 A/g, respectively). P=

ME

0.8. A device previously cycled, or “primed,” to 0.5 V for 10 000 times at 25 A/g retains 93% capacitance when cycled 10 000 times under 0.8 V at 10 A/g. Strikingly, a device cycled at 0.8 V with no previous priming at 0.5 V retains only 87% capacitance 29457


Research Article


gently and rapidly mixed and set aside for 12 h. The PANi is purified through vacuum filtration. Transmission electron micrographs were collected using an FEI Tecnai G2 Spirit TEM, with 120 kV accelerating voltage. The sample was drop-cast from a dispersion of lyophilized PANi dispersed in hexafluoroisopropanol (HFIP) onto a carbon coated Cu grid. Scanning electron micrographs of PANi cast from an aqueous dispersion onto aluminum were collected with an FEI Nova 230. UV−vis spectra were collected with an OceanOptics HL2000-FHSA light source on an STS Microspectrometer. Lyophilization of PANi. Wet, filtered PANi is poured into small beaker and submerged in liquid nitrogen until frozen. The frozen PANi is placed into a vessel with a vacuum seal which is then insulated. A vacuum is pulled on the round mouthed beaker through a trap containing acetone and dry ice. Dry ice is periodically resupplied in order to keep the trap at −78 °C. The vacuum is held until only dry PANi powder remains in the vessel. Dispersion of PANi in PLA Binder Solution. A 2 wt % PLA in benzene mixture is mechanically agitated at room temperature over 2 days until fully solubilized. Lyophilized PANi (2 g) is dispersed in 100 mL of 2 wt % PLA in benzene solution. PLA is also soluble in chloroform, nitromethane, carbon disulfide, benzene, hexafluoroisopropanol, trifluoroethanol, octafluoropentanol, and tetrafluoropropanol. The fluorinated alcohols benefit from low surface tensions for facile drop-casting of homogeneous thin films. Among these, trifluoroethanol is the least cost-prohibitive. For the purposes of this work, an inexpensive solvent, benzene, was chosen to demonstrate the versatility of this technique. The dispersion is mechanically agitated over 12 hno destructive techniques (such as sonication) were used in order to preserve the integrity of the nanofibers. In cases where activated carbon is used as a control, a similar process for dispersion is followed, where activated carbon (Norit DLC Supra, 3 g) is added to 100 mL of 2 wt % PLA in benzene. Electrode Preparation. The solutions described in the previous section are drop-cast onto a piece of hard carbon paper (hcp). The hcp is suspended in air during this process in order to mitigate wicking of the solution. The droplet is dried under 2 scfm N2 bubbled through water at 35 °C, forming a mechanically robust web of PLA/PANi. Electrochemical Characterization. To form the working electrode for three-electrode cyclic voltammetry, platinum foil is affixed to the back of a PANi on hcp electrode using polyimide tape. A platinum mesh current collector, an Ag/AgCl reference electrode, and the working electrode are submerged in a 10 mL beaker containing 1 M H2SO4. The fabrication process for a two-electrode device is detailed in Figure 3. In both two- and three-electrode geometries, the leads are connected to a Bio-Logic VMP3 potentiostat for electrochemical characterization. For open circuit potential measurements, no current or potential is applied as the potential between the working and reference electrodes is measured by the potentiostat. All calculations are made based on the mass of the active material including PLA binder (averaged across both electrodes for twoelectrode geometry). Resistance values and I−V curves were taken using a Keithley 2450 SourceMeter with a convertible 2/4 point probe station built in-house by author Y.L., where two gold probes are held at a 1 mm distance from each other with equal pressure on a sample (Figure S6). XPS analysis was carried out using an Al Kα X-ray source from Physical Electronics, 5000 VersaProbe II Scanning Microprobe system.

Figure 8. Performance metrics for devices studied: (i) 0.8 V PANi EC, (ii) primed 0.8 V PANi EC, (iii) 0.5 V PANi EC, and (iv) 0.8 V activated carbon EC. (a) Ragone plot of all devices. (b) Superimposed galvanostatic charge−discharge curves for devices ii and iv at 0.5 and 1 A/g.

By comparing curves iii and iv in Figure 8a, it is immediately clear that a PANi EC, even when constrained to 0.5 V, has a vastly increased energy density over that of an activated carbon device fabricated without conductive additives. Activated carbon accepts ions into the pores at the surface of rigid grains of active material, forming a double layer. It has a moderate amount of acidic sites on the surface of the particles which also contribute a certain amount of pseudocapacitance.58 However, in protonically dopable polymers such as polyaniline, the individual chains themselves can be protonated and deprotonated with ease and without destruction of the electronic transport pathway, leading to a high energy density at surprisingly low voltage windows.21,22

■

CONCLUSIONS Our work addresses the misconception of instability associated with symmetric PANi ECs, and demonstrates how to vastly enhance the cycling stability of two-electrode systems. By carefully focusing on the electrochemistry at each electrode as well as the mechanism of charge storage, a PANi EC can undergo 150 000 cycles at 25 A/g with 98% capacitance retention; this is comparable to aqueous EDLCs, while exhibiting a specific energy 64 times greater than an identical EC made with activated carbon without the use of any conductive additives. A further tripling of the specific energy is achieved by expanding the voltage window to 0.8 V. Normally, this expansion leads to a quick loss of capacitance, retaining only 87% of the initial capacitance value over 10 000 cycles at 10 A/ g. However, by first cycling at a lower window of 0.5 V, i.e., priming the device, a capacitance retention of 93% over 10 000 cycles at 10 A/g is achieved. Specific energy of 0.8 V devices reaches 33.39 Wh/kg at 600 W/kg, further showcasing PANi as a promising candidate for advanced aqueous electrochemical capacitors.

■

■

ASSOCIATED CONTENT

S Supporting Information *

EXPERIMENTAL METHODS

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsami.6b09779. Three-electrode cyclic voltammograms of PANi in various electrolytes and for different voltage windows; hard carbon paper cyclic voltammograms and charge− discharge curves; deconvolutions of high-resolution XPS spectra of C1s peaks of PANi; SEM of hard carbon paper; schematic of I−V curve measurement (PDF)

Preparation of PANi. PANi is synthesized using a dimer initiated solution oxidative polymerization technique described in detail by Tran et al.59 Aniline (16 mL, 0.175 mol) is added to 800 mL of 1 M ptoluenesulfonic acid (p-TSA), and ammonium peroxydisulfate (APS, 10.02 g, 0.044 mol) is dissolved in 800 mL of 1 M p-TSA. N-Phenyl1,4-phenylenediamine (0.32 g, 1.2 mmol) is dissolved in 100 mL of methanol. The methanol solution containing the aniline dimer is combined with aniline aqueous solution. Next, APS solution is combined with monomer/dimer solution, and the entire solution is 29458


Research Article


■

(19) Snook, G. A.; Kao, P.; Best, A. S. Conducting-Polymer-Based Supercapacitor Devices and Electrodes. J. Power Sources 2011, 196, 1− 12. (20) Huang, Y.; Zhu, M.; Pei, Z.; Huang, Y.; Geng, H.; Zhi, C. Extremely Stable Polypyrrole Achieved Via Molecular Ordering for Highly Flexible Supercapacitors. ACS Appl. Mater. Interfaces 2016, 8, 2435−2440. (21) Barsukov, V.; Chivikov, S. New Model for the Mechanism of Current-Producing Processes in Polyaniline-Type Conductive Polymers. Mater. Sci. Forum 1995, 191, 257−260. (22) Barsukov, V.; Chivikov, S. The “Capacitor” Concept of the Current-Producing Process Mechanism in Polyaniline-Type Conducting Polymers. Electrochim. Acta 1996, 41, 1773−1779. (23) Khomenko, V.; Frackowiak, E.; Béguin, F. Determination of the Specific Capacitance of Conducting Polymer/Nanotubes Composite Electrodes Using Different Cell Configurations. Electrochim. Acta 2005, 50, 2499−2506. (24) Bunz, U. H. F. Breath Figures as a Dynamic Templating Method for Polymers and Nanomaterials. Adv. Mater. 2006, 18, 973−989. (25) Zhang, A.; Bai, H.; Li, L. Breath Figure: A Nature-Inspired Preparation Method for Ordered Porous Films. Chem. Rev. 2015, 115, 9801−9868. (26) Ma, C.-Y.; Zhong, Y.-W.; Li, J.; Chen, C.-K.; Gong, J.-L.; Xie, S.Y.; Li, L.; Ma, Z. Patterned Carbon Nanotubes with Adjustable Array: A Functional Breath Figure Approach. Chem. Mater. 2010, 22, 2367− 2374. (27) Böker, A.; Lin, Y.; Chiapperini, K.; Horowitz, R.; Thompson, M.; Carreon, V.; Xu, T.; Abetz, C.; Skaff, H.; Dinsmore, A. D.; Emrick, T.; Russell, T. P. Hierarchical Nanoparticle Assemblies Formed by Decorating Breath Figures. Nat. Mater. 2004, 3, 302−306. (28) Gogotsi, Y.; Simon, P. True Performance Metrics in Electrochemical Energy Storage. Science 2011, 334, 917−918. (29) Williams, G.; McMurray, H. N. Factors Affecting Acid-Base Stability of the Interface between Polyaniline Emeraldine Salt and Oxide Covered Metal. Electrochem. Solid-State Lett. 2005, 8, B42−B45. (30) Barsukov, V. Z.; Khomenko, V. G.; Chivikov, S. V.; Barsukov, I. V.; Motronyuk, T. I. On the Faradaic and Non-Faradaic Mechanisms of Electrochemical Processes in Conducting Polymers and Some Other Reversible Systems with Solid-Phase Reagents. Electrochim. Acta 2001, 46, 4083−4094. (31) Khomenko, V.; Raymundo-Piñero, E.; Frąckowiak, E.; Béguin, F. High-Voltage Asymmetric Supercapacitors Operating in Aqueous Electrolyte. Appl. Phys. A: Mater. Sci. Process. 2006, 82, 567−573. (32) Gu, J.; Kan, S.; Shen, Q.; Kan, J. Effects of Sulfanilic Acid and Anthranilic Acid on Electrochemical Stability of Polyaniline. Int. J. Electrochem. Sci. 2014, 9, 6858−6869. (33) Arsov, D. L.; Plieth, W.; Koßmehl, G. Electrochemical and Raman Spectroscopic Study of Polyaniline; Influence of the Potential on the Degradation of Polyaniline. J. Solid State Electrochem. 1998, 2, 355−361. (34) Geniès, E. M.; Lapkowski, M.; Penneau, J. F. Cyclic Voltammetry of Polyaniline: Interpretation of the Middle Peak. J. Electroanal. Chem. Interfacial Electrochem. 1988, 249, 97−107. (35) Gupta, V.; Miura, N. High Performance Electrochemical Supercapacitor from Electrochemically Synthesized Nanostructured Polyaniline. Mater. Lett. 2006, 60, 1466−1469. (36) Sivakkumar, S. R.; Kim, W. J.; Choi, J.-A.; MacFarlane, D. R.; Forsyth, M.; Kim, D.-W. Electrochemical Performance of Polyaniline Nanofibres and Polyaniline/Multi-Walled Carbon Nanotube Composite as an Electrode Material for Aqueous Redox Supercapacitors. J. Power Sources 2007, 171, 1062−1068. (37) Gupta, V.; Miura, N. Polyaniline/Single-Wall Carbon Nanotube (PANI/SWCNT) Composites for High Performance Supercapacitors. Electrochim. Acta 2006, 52, 1721−1726. (38) Xu, W.; Mu, B.; Zhang, W.; Wang, A. Facile Fabrication of WellDefined Polyaniline Microtubes Derived from Natural Kapok Fiber for Supercapacitor with Long-Term Cycling Stability. RSC Adv. 2016, 6, 68302.

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS L.M.S. gratefully acknowledges the financial support of the U.S. Department of Education through a GAANN Fellowship, PR/ Award No. P200A150105.

■

REFERENCES

(1) International Energy Outlook 2016; DOE/EIA-0484(2016); U.S. Department of Energy: 2016. (2) Simon, P.; Gogotsi, Y.; Dunn, B. Where Do Batteries End and Supercapacitors Begin? Science 2014, 343, 1210−1211. (3) Frąckowiak, E.; Béguin, F. Carbon Materials for the Electrochemical Storage of Energy in Capacitors. Carbon 2001, 39, 937−950. (4) Devaraj, S.; Munichandraiah, N. Effect of Crystallographic Structure of MnO2 on Its Electrochemical Capacitance Properties. J. Phys. Chem. C 2008, 112, 4406−4417. (5) Julien, C.; Ziolkiewicz, S.; Lemal, M.; Massot, M. Synthesis, Structure and Electrochemistry of LiMn2‑yAlyO4 Prepared by a WetChemistry Method. J. Mater. Chem. 2001, 11, 1837−1842. (6) Rasmussen, S. C. Electrically Conducting Plastics: Revising the History of Conjugated Organic Polymers. In 100+ Years of Plastics. Leo Baekeland and Beyond; Strom, E. T., Rasmussen, S. C., Eds.; ACS Symposium Series 1080; American Chemical Society: Washington, DC, 2011; pp 147−163. DOI: 10.1021/bk-2011-1080.ch010. (7) Pople, J. A.; Walmsley, S. H. Electronic States of Long Polyenes with Alternating Bond Lengths. Trans. Faraday Soc. 1962, 58, 441− 448. (8) Bolto, B. A.; McNeill, R.; Weiss, D. E. Electronic Conduction in Polymers. Aust. J. Chem. 1963, 16, 1090−1103. (9) Shirakawa, H.; Louis, E. J.; MacDiarmid, A. G.; Chiang, C. K.; Heeger, A. J. Synthesis of Electrically Conducting Organic Polymers: Halogen Derivatives of Polyacetylene, (CH)x. J. Chem. Soc., Chem. Commun. 1977, 578−580. (10) Curgul, S.; Van Vliet, K. J.; Rutledge, G. C. Molecular Dynamics Simulation of Size-Dependent Structural and Thermal Properties of Polymer Nanofibers. Macromolecules 2007, 40, 8483−8489. (11) Zhang, W.; Charleux, B.; Cassagnau, P. Viscoelastic Properties of Water Suspensions of Polymer Nanofibers Synthesized Via RaftMediated Emulsion Polymerization. Macromolecules 2012, 45, 5273− 5280. (12) Khomenko, V.; Raymundo-Piñero, E.; Béguin, F. Optimisation of an Asymmetric Manganese Oxide/Activated Carbon Capacitor Working at 2V in Aqueous Medium. J. Power Sources 2006, 153, 183− 190. (13) Ryu, K. S.; Kim, K. M.; Park, N.-G.; Park, Y. J.; Chang, S. H. Symmetric Redox Supercapacitor with Conducting Polyaniline Electrodes. J. Power Sources 2002, 103, 305−309. (14) Laforgue, A.; Simon, P.; Sarrazin, C.; Fauvarque, J.-F. Polythiophene-Based Supercapacitors. J. Power Sources 1999, 80, 142−148. (15) Di Fabio, A.; Giorgi, A.; Mastragostino, M.; Soavi, F. CarbonPoly(3-Methylthiophene) Hybrid Supercapacitors. J. Electrochem. Soc. 2001, 148, A845−A850. (16) Otero, T. F.; Martinez, J. G. Structural and Biomimetic Chemical Kinetics: Kinetic Magnitudes Include Structural Information. Adv. Funct. Mater. 2013, 23, 404−416. (17) Otero, T. F.; Ariza, M. J. Revisiting the Electrochemical and Polymeric Behavior of a Polypyrrole Free-Standing Electrode in Aqueous Solution. J. Phys. Chem. B 2003, 107, 13954−13961. (18) Béguin, F.; Presser, V.; Balducci, A.; Frąckowiak, E. Carbons and Electrolytes for Advanced Supercapacitors. Adv. Mater. 2014, 26, 2219−2251. 29459


Research Article


Produced Via Rapid Initiated Polymerization. Macromolecules 2008, 41, 7405−7410.

(39) Chang, W.-M.; Wang, C.-C.; Chen, C.-Y. Plasma-Induced Polyaniline Grafted on Carbon Nanotube-Embedded Carbon Nanofibers for High-Performance Supercapacitors. Electrochim. Acta 2016, 212, 130−140. (40) Rudge, A.; Davey, J.; Raistrick, I.; Gottesfeld, S.; Ferraris, J. P. Conducting Polymers as Active Materials in Electrochemical Capacitors. J. Power Sources 1994, 47, 89−107. (41) Astakhov, M. V.; Galimzyanov, R. R.; Klimont, A. A.; Krechetov, I. S.; Kundu, M.; Stakhanova, S. V. Improved Symmetric Supercapacitive Performance of Binder-Free PANI/Carbon Fiber Composites. Curr. Nanosci. 2016, 12, 83−89. (42) Khalid, M.; Tumelero, M. A.; Pasa, A. A. Asymmetric and Symmetric Solid-State Supercapacitors Based on 3D Interconnected Polyaniline-Carbon Nanotube Framework. RSC Adv. 2015, 5, 62033− 62039. (43) Pourbaix, M. Atlas of Electrochemical Equilibria in Aqueous Solutions; Pergamon Press: Oxford, U.K., 1966. (44) Varcoe, J. R.; Atanassov, P.; Dekel, D. R.; Herring, A. M.; Hickner, M. A.; Kohl, P. A.; Kucernak, A. R.; Mustain, W. E.; Nijmeijer, K.; Scott, K.; Xu, T.; Zhuang, L. Anion-Exchange Membranes in Electrochemical Energy Systems. Energy Environ. Sci. 2014, 7, 3135−3191. (45) Luo, J.-Y.; Cui, W.-J.; He, P.; Xia, Y.-Y. Raising the Cycling Stability of Aqueous Lithium-Ion Batteries by Eliminating Oxygen in the Electrolyte. Nat. Chem. 2010, 2, 760−765. (46) Maletin, Y.; Novak, P.; Shembel, E.; Izotov, V.; Strizhakova, N.; Mironova, A.; Danilin, V.; Podmogilny, S. Matching the Nanoporous Carbon Electrodes and Organic Electrolytes in Double Layer Capacitors. Appl. Phys. A: Mater. Sci. Process. 2006, 82, 653−657. (47) Khomenko, V.; Raymundo-Piñero, E.; Béguin, F. A New Type of High Energy Asymmetric Capacitor with Nanoporous Carbon Electrodes in Aqueous Electrolyte. J. Power Sources 2010, 195, 4234− 4241. (48) Prasad, K. R.; Munichandraiah, N. Fabrication and Evaluation of 450 F Electrochemical Redox Supercapacitors Using Inexpensive and High-Performance, Polyaniline Coated, Stainless-Steel Electrodes. J. Power Sources 2002, 112, 443−451. (49) Dai, Z.; Peng, C.; Chae, J. H.; Ng, K. C.; Chen, G. Z. Cell Voltage Versus Electrode Potential Range in Aqueous Supercapacitors. Sci. Rep. 2015, 5, 9854. (50) Rubinson, J. F.; Kayinamura, Y. P. Charge Transport in Conducting Polymers: Insights from Impedance Spectroscopy. Chem. Soc. Rev. 2009, 38, 3339−3347. (51) Taberna, P.-L.; Simon, P. Electrochemical Techniques. In Supercapacitors; Wiley-VCH: 2013; pp 111−130. (52) Schweiger, H.; Multerer, M.; Schweizer-Berberich, M.; Gores, H. Optimization of Cycling Behavior of Lithium Ion Cells at 60 °C by Additives for Electrolytes Based on Lithium Bis[1,2-Oxalato(2-)-O,O′] Borate. Int. J. Electrochem. Sci. 2008, 3, 427−443. (53) Irvin, J. A.; Irvin, D. J.; Stenger-Smith, J. D. Electroactive Polymers for Batteries and Supercapacitors. In Conjugated Polymers; CRC Press: 2006; pp 9-1−9-29. (54) Kurra, N.; Wang, R.; Alshareef, H. N. All Conducting Polymer Electrodes for Asymmetric Solid-State Supercapacitors. J. Mater. Chem. A 2015, 3, 7368−7374. (55) Jurewicz, K.; Frąckowiak, E.; Béguin, F. Towards the Mechanism of Electrochemical Hydrogen Storage in Nanostructured Carbon Materials. Appl. Phys. A: Mater. Sci. Process. 2004, 78, 981− 987. (56) Kumar, S. N.; Gaillard, F.; Bouyssoux, G.; Sartre, A. HighResolution XPS Studies of Electrochemically Synthesized Conducting Polyaniline Films. Synth. Met. 1990, 36, 111−127. (57) Briggs, D. Introduction. In Surface Analysis of Polymers by XPS and Static SIMS; Cambridge University Press: 1998; Chapter 1. (58) Frąckowiak, E. Carbon Materials for Supercapacitor Application. Phys. Chem. Chem. Phys. 2007, 9, 1774−1785. (59) Tran, H. D.; Norris, I.; D’Arcy, J. M.; Tsang, H.; Wang, Y.; Mattes, B. R.; Kaner, R. B. Substituted Polyaniline Nanofibers 29460


Enhancing Cycling Stability of Aqueous ... - ACS Publications

Recommend Documents